0% found this document useful (0 votes)
581 views

DSP Formula

This document provides formulas and properties for various signals and systems in digital signal processing (DSP). It includes formulas for even and odd functions, roots of polynomials, complex numbers, trigonometric functions, frequencies and periods, integral transforms, convolution, correlation, random signals, Fourier series, Fourier transforms, z-transforms, and unit impulse and step functions. The document is a reference table of common DSP formulas for exam preparation.

Uploaded by

dangtran_nam
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
581 views

DSP Formula

This document provides formulas and properties for various signals and systems in digital signal processing (DSP). It includes formulas for even and odd functions, roots of polynomials, complex numbers, trigonometric functions, frequencies and periods, integral transforms, convolution, correlation, random signals, Fourier series, Fourier transforms, z-transforms, and unit impulse and step functions. The document is a reference table of common DSP formulas for exam preparation.

Uploaded by

dangtran_nam
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

T-61.3010 DSP 2006 (F) 1/2 v. 1.

04, 2006-04-25

T-61.3010 DSP Table of formulas, spring 2006


Disclaimer! Notations, e.g. ω or Ω, may vary from book to book, or from exam paper to other.

Even and odd functions: Frequencies, angular frequencies, periods:


Even{x(t)} = 0.5 · [x(t) + x(−t)] Here fs is sampling frequency (also fT later)
Odd{x(t)} = 0.5 · [x(t) − x(−t)] Frequency
f , [f ] = Hz = 1/s
Roots of second-order polynomial:
Angular frequency
ax2 + bx +√
c=0
Ω = 2πf = 2π/T , [Ω] = rad/s
x = (−b ± b2 − 4ac)/(2a)
Normalized angular frequency
Complex √ numbers: ω = 2πΩ/Ωs = 2πf /fs , [ω] = rad/sample
i ≡ j = −1 = − 1j Normalized frequency in Matlab
z=p x + jy = r ejθ fMAT LAB = 2f /fs , [fMAT LAB ] = 1/sample
r = x2 + y 2 , θ = arctan(y/x) Integral transform properties
x = r cos(θ), y = r sin(θ) Here all integral transforms share some basic properties.
ejθ = cos(θ)
√ + j sin(θ) Examples given with CTFT, x[n] ↔ X(ejω ), x1 [n] ↔
√ N √
N
z = r ejθ = | N r| ej(θ+2πk)/N , k = 0, 1, 2, . . . , N − 1 X1 (ejω ), and x2 [n] ↔ X2 (ejω ) are time-domain signals with
corresponding transform-domain spectra. a and b are cons-
Trigonometric functions:
tants.
sinc(θ) = sin(πθ)/(πθ)
Linearity. All transforms are linear.
sin(θ)/θ → 1, when θ → 0; sinc(θ) → 1, when θ → 0
ax1 [n] + bx2 [n] ↔ aX1 (ejω ) + bX2 (ejω )
cos(θ) = 0.5ejθ + 0.5e−jθ
Time-shifting. There is a kernel term in transform, e.g.,
sin(θ) = 0.5je−jθ − 0.5jejθ
x[n − k] ↔ e−jkω X(ejω )
cos2 (θ) + sin2 (θ) = 1
Frequency-shifting. There is a kernel term in signal e.g.,
θ 0 π/6 π/4 π/3
√ √ ejωk n x[n] ↔ X(ej(ω−ωk ) )
sin(θ) 0 0.5
√ √ 2/2 3/2 Conjugate symmetry.
cos(θ) 1 3/2 2/2 0.5 x∗ [n] ↔ X ∗ (e−jω )
θ π/2 3π/4√ π −π/2 If x[n] ∈ R, then
sin(θ) 1 2/2
√ 0 −1 X(ejω ) = X ∗ (e−jω )
cos(θ) 0 − 2/2 −1 0 |X(ejω )| = |X(e−jω )|
√ √
π ≈ 3.1416, 3/2 ≈ 0.8660, 2/2 ≈ 0.7071 ∠X(ejω ) = −∠X(e−jω )
If x[n] ∈ R and even, then X(ejω ) ∈ R and even.
Geometric series:
P+∞ n 1 If x[n] ∈ R and odd, then X(ejω ) purely ∈ C and odd.
n=0 a = 1 − a , |a| < 1 Time reversal. Transform variable is reversed, e.g.,
PN n 1 − aN +1 , |a| < 1 x[−n] ↔ X(e−jω )
n=0 a = 1−a
Differentiation. In time and frequency domain, e.g.,
Continuous-time unit step and unit impulse func- x[n] − x[n − 1] ↔ (1 − e−jω )X(ejω )
tions: ( nx[n] ↔ j dω d
X(ejω )
1, t > 0 Duality. Convolution property: convolution in time domain
µ(t) =
0, t < 0 corresponds multiplication in transform domain
δ (t) = dtd
µ∆ (t), δ(t) = lim∆→0 δ∆ (t) x1 [n] ⊛ x2 [n] ↔ X1 (ejω ) · X2 (ejω )
R∆∞
δ(t) dt =1 and multipication R property: vice versa
R−∞
∞ x1 [n] · x2 ↔ 2π1
X1 (ejθ )X2 (ej(ω−θ) ) dθ
−∞
δ(t − t0 )x(t) dt = x(t0 ) 2π
R Parseval’s relation. Energy in signal and spectral compo-
In DSP notation 2πδ(t) is computed 2π δ(t) · 1 dt = 2π,
nents:
when t = 0, and = 0 elsewhere. 1
|x[n]|2 = 2π jω 2
P R
2π |X(e )| dω
Discrete-time unit impulse and unit step functions:
( Integral transforms
1, n = 0 Definitions given in first two lines of each type. Some com-
δ[n] =
0, n 6= 0 mon pairs as well as properties are listed. See math reference
(
1, n ≥ 0 book for complete tables.
µ[n] =
0, n < 0 Fourier-series
P∞ of continuous-time periodic signals:
x(t) = k=−∞ ak ejkΩ0 t
Convolution
ak = T1 T x(t) e−jkΩ0 t dt
R
Convolution is commutative, associative and distributive.
R∞
y(t) = h(t) ⊛ x(t) = −∞ h(τ )x(t − τ ) dτ x(t − t0 ) ↔ ak ejkΩ0 t0
jMΩ0 t
y[n] = h[n] ⊛ x[n] = +∞
P Re x(t) ↔ ak−M
k=−∞ h[k]x[n − k] x (τ )xb (t −
T a Pτ ) dτ ↔ T ak bk
Correlation: xa (t)xb (t) ↔ l al bk−l
P+∞ d
rxy [l] = n=−∞ x[n]y[n − l] = x[l] ⊛ y[−l] dt x(t) ↔ jkΩ0 ak
P+∞
rxx [l] = n=−∞ x[n]x[n − l] Fourier-series of discrete-time periodic sequences:
x[n] = k=hN i ak ejkω0 n , x[n] periodic with N0
P
Mean and variance of random signal:
ak = N1 n=hN i x[n] e−jkω0 n , ak periodic with N0
R P
mX =RE[X] = xpX (x)dx
2
σX = (x − mX )2 pX (x)dx = E[X 2 ] − m2X x[n − M ] ↔ ak ejkω0 M
T-61.3010 DSP 2006 (F) 2/2 v. 1.04, 2006-04-25

ejMω0 n x[n] ↔ ak−M δ[n] ↔ 1, ROC all z


δ[n − k] ↔ z −k , all z, except 0 (k > 0) or ∞ (k < 0)
Continuous-time Fourier-transform (CTFT):
1
R∞ µ[n] ↔ 1−z1 −1 , |z| > 1
x(t) = 2π X(jΩ) ejΩt dΩ
R −∞
∞ −jΩt −µ[−n − 1] ↔ 1−z1 −1 , |z| < 1
X(jΩ) = −∞ x(t) e dt 1
jΩtk an µ[n] ↔ 1−az −1 , |z| > |a|
x(t − tk ) ↔ e X(jΩ)
az
nan µ[n] ↔ (1−az
−1
ejΩk t x(t) ↔ X(j(Ω − Ωk )) −1 )2 , |z| > |a|
xa (t) ⊛ xb (t) ↔ Xa (jΩ)Xb (jΩ) n 1
(n + 1)a µ[n] ↔ (1−az−1 )2 , |z| > |a|
1
xa (t)xb (t) ↔ 2π Xa (jΩ) ⊛ Xb (jΩ) 1−r cos(ω0 )z −1
d rn cos(ω0 n)µ[n] ↔ 1−2r cos(ω0 )z −1 +r 2 z −2 , |z| > |r|
dt x(t) ↔ jΩX(jΩ)
r sin(ω0 )z −1
tx(t) ↔ j dΩ d
X(jΩ) rn sin(ω0 n)µ[n] ↔ 1−2r cos(ω0 )z −1 +r 2 z −2 , |z| > |r|
jΩ0 t
e ↔ 2πδ(Ω − Ω0 ) LTI filter
cos(Ω0 t) ↔ π[δ(Ω − Ω0 ) + δ(Ω + Ω0 )] Panalysis
Stability h |h[n]| < ∞; unit cirle belongs to ROC
sin(Ω0 t) ↔ jπ[δ(Ω + Ω0 ) − δ(Ω − Ω0 )] Causality h[n] = 0, n < 0; P
∞ belongs to ROC
x(t) = ( 1 ↔ 2πδ(Ω) n
Unit step response s[n] = k=−∞ h[k]
1, |t| < T1 Causal transfer function of order max{M, N }:
x(t) = ↔ 2 sin(ΩT

1)
M M
0, |t| > T1
P Q
bm z −m (1−dm z −1 )
H(z) = K · Pm=0
N = G· Qm=1
N
n=0 an z n=1 (1−pn z
( −n −1 )

sin(W t) 1, |Ω| < W Zeros: B(z) = 0; Poles: A(z) = 0; where H(z) = B(z)/A(z)
πt ↔ X(jΩ) =
0, |Ω| > W Frequency, magnitude/amplitude, phase response, z ← ejω

δ(t) ↔ 1 H(ejω ) = |H(ejω )| ej∠H(e )
δ(t − tk ) ↔ ejΩtk jω
H(e ) = H(z)|z=ejω
1
e−at µ(t) ↔ a+jΩ , where Real{a} > 0 H[k] = H(ejω )|ω=2πk/N
d
Group delay τ (ω) = − dω ∠H(ejω )
Discrete-time Fourier-transform (DTFT):
1 jω Important transform pairs and properties:
) ejωn dω
R
x[n] = 2π 2π∞X(e
X(e ) = n=−∞ x[n] e−jωn , X(ejω ) periodic 2π

P a δ[n − k] ↔ a e−jkω ↔ a z −k
an µ[n] ↔ 1/[1 − a e−jω ] ↔ 1/[1 − a z −1 ]
x[n − k] ↔ e−jkω X(ejω )
a x[n − k] ↔ a e−jkω X(ejω ) ↔ a z −k X(z)
ejωk n x[n] ↔ X(ej(ω−ωk ) )
y[n] = h[n] ⊛ x[n] ↔ Y (z) = H(z) · X(z)
x1 [n] ⊛ x2 [n] ↔RX1 (ejω ) · X2 (ejω )
1 jθ j(ω−θ) rectangular ↔ sinc, sinc ↔ rectangular
x1 [n] · x2 ↔ 2π 2π X1 (e )X2 (e ) dθ
d jω LTI filter design (synthesis)
nx[n] ↔ j dωP X(e )
ejω0 n ↔ 2π P l δ(ω − ω0 − 2πl)
Bilinear transform H(z) = H(s)|s and prewarping
cos(ω0 n) ↔ π Pl [δ(ω − ω0 − 2πl) + δ(ω + ω0 − 2πl) s = k · (1 − z −1 )/(1 + z −1 ), k = 1 or k = 2/T = 2fT
sin(ω0 n) ↔ jπ P l [δ(ω + ω0 − 2πl) − δ(ω − ω0 − 2πl) Ωprewarp,c = k · tan(ωc /2), k = 1 or k = 2/T = 2fT
x[n] = ( 1 ↔ 2π l δ(ω − 2πl) Spectral transformations, ω̂c desired cut-off
1, |n| ≤ N1 LP-LP z −1 = (ẑ −1 − α)/(1 − αẑ −1 ), where
x[n] = ↔ sin(ω(N 1 +0.5))
sin(ω/2) α = sin(0.5(ωc − ω̂c ))/ sin(0.5(ωc + ω̂c ))
0, |n| > N1
sin(W n)
LP-HP z −1 = −(ẑ −1 + α)/(1 + αẑ −1 ), where
W Wn
πn = πsinc(
( π ) ↔ ... α = − cos(0.5(ωc + ω̂c ))/ cos(0.5(ωc − ω̂c ))
1, 0 ≤ |ω| ≤ W Windowed(Fourier series method
. . . X(ejω ) = 1, |ω| < ωc
0, W < |ω| ≤ π H(ejω ) = ↔ ...
δ[n] ↔ 1 0, |ω| ≥ ωc
δ[n − k] ↔ e−jkω . . . h[n] = sin(ω c n)
= ωπc sinc( ωπc n )
πn
an µ[n] ↔ 1−ae1−jω , |a| < 1 hF IR [n] = hideal [n]
1
R π · w[n]
Discrete Fourier-transform (DFT): HF IR (ejω ) = 2π jθ
−π Hideal (e )W (e
j(ω−θ)
) dθ
Connection to DTFT: X[k] = X(ejω )|ω=2πk/N Fixed window functions, order N = 2M , −M ≤ n ≤ M :
WN = e−j2π/N Rectangular w[n] = 1
x[n] = N1
PN −1 −kn Hamming w[n] = 0.54 + 0.46 cos((2πn)/(2M  ))
k=0 X[k]WN , 0≤k ≤N −1
PN −1 kn Hann w[n] = 0.5 · 1 + cos((2πn)/(2M ))
X[k] = n=0 x[n]WN , 0 ≤ k ≤ N − 1
Blackman w[n] = 0.42 + 0.5 cos((2πn)/(2M )) +
Laplace transform: 0.08 cos((4πn)/(2M ))
Convergence with a certain ROC (region of convergence). Bartlett w[n] = 1 − (|n|/M )
Connection to continuous-time Fourier-transform: s = jΩ
1
R σ+j∞ Multirate systems
x(t) = 2πj X(s)est ds
R ∞ σ−j∞ −st Upsampling with factor L, ↑ L
X(s) = −∞ x(t) e dt (
x[n/L], n = 0, ±L, ±2L, . . .
z-transform: xu [n] =
xu [n] = 0, otherwise
Convergence with a certain ROC (region of convergence).
ConnectionH to discrete-time Fourier-transform: z = ejω Xu (z) = X(z L ), Xu (ejω ) = X(ejωL )
1
x[n] = 2πj X(z)z n−1 dz Downsampling with factor M , ↓ M
P∞
X(z) = n=−∞ x[n]z −n xd [n] = x[nM ]
PM−1 −k
x[n − k] ↔ z −k X(z) Xd (z) = (1/M ) k=0 X(z 1/M WM ),

PM−1 j(ω−2πk)/M
x1 [n] ⊛ x2 [n] ↔ X1 (z) · X2 (z) Xd (e ) = (1/M ) k=0 X(e )

You might also like